[13:59] <mattgriffin> Daviey: hiya. can you send me and stewart the url for the hangout?
[13:59]  * stewart introduces self: Stewart Smith, Director of Software Architecture at Percona.
[13:59]  * mattgriffin is Matt Griffin, Director of Product Management at Percona
[14:00] <Daviey> mattgriffin: i think you have it?
[14:00] <Daviey> waiitng to start..
[14:01] <Daviey> starting...
[14:02] <jamespage> can people see the stream yet?
[14:02] <mattgriffin> Daviey: don't think we're on the air yet
[14:03] <stewart> i can see everyone at least :)
[14:03]  * stewart notes everybody has cool lines of text and Logos except me
[14:04] <Daviey> we are online..
[14:05] <stewart> or this morning if you're me :)
[14:09] <jamespage> https://wiki.ubuntu.com/ServerTeam/OpenStackHA
[14:13] <Daviey> (If anyone else thinks they should be in he inner circle, please shout)
[14:15]  * med_ is just watching
[14:30] <Daviey> stewart: can you remind me of your LP id?
[14:30] <stewart> Daviey, stewart
[14:30] <Daviey> ta
[14:30] <Daviey> (thanks was harder than i expected)
[14:32] <stewart> :)
[14:51] <roaksoax> Link: http://blog.clusterlabs.org/
[14:59] <med_> thanks percona.
[14:59] <stewart> np. thanks all
[15:03] <Daviey> Who is here for the mongdo session?
[15:04] <Daviey> rbasak: joining us?
[15:05] <rbasak> Yes
[15:06] <rbasak> Having trouble getting the untypeable link across to my other machine
[15:07] <Daviey> rbasak: you could use a url shortenr :)
[15:08] <jamespage> med_, still here?
[15:08] <med_> yep
[15:09] <med_> I see starting soon
[15:09] <jamespage> refersh
[15:09] <Daviey> started
[15:09] <rbasak> Apparently lower third doesn't work on ARM :(
[15:10] <Daviey> Anyone else want to be in the inner circle?
[15:10] <med_> live now
[15:10] <med_> Daviey, Hyde Park?
[15:10] <Daviey> med_: ^ ?
[15:10] <Daviey> med_: Yes, i was busy hunting deer.
[15:11] <yolanda> happybase is also supported by ceilometer
[15:12] <Daviey> yolanda: interesting
[15:15] <zyga-uds> loud typing
[15:17] <zyga-uds> loud typing
[15:17] <zyga-uds> loud typing
[15:20] <Daviey> zyga: noted, thanks
[15:21] <med_> j a m e s p a g e   can't efficiently mute and talk and take notes at the same time... so we tolerate the typing noise.
[15:22] <med_> youtube just froze
[15:22] <med_> and back
[15:23]  * zyga didn't indent to say 'loud typing' three times, the irc web widget is borken,
[15:23] <zyga> I was actually interested in the mongo patches, where are they?
[15:24] <rbasak> zyga: http://bazaar.launchpad.net/~james-page/ubuntu/saucy/mongodb/merge-2.4.3/files/head:/debian/patches/ for James' latest work in progress
[15:24] <jamespage> any questions from channel
[15:24] <med_> does MRE limit who can submit that?
[15:25] <rbasak> zyga: http://bazaar.launchpad.net/~ubuntu-branches/ubuntu/saucy/mongodb/saucy/files/head:/debian/patches/ for the latest in Saucy
[15:25] <med_> ie does it require someone have uploader status?
[15:26] <yolanda> hbase, ceilometer team chose mongo because was a customer requirement only
[15:26] <med_> not sure how much lag between stream and live... seems pretty laggy based on the comments I'm hearing
[15:26] <yolanda> but they said in ODS ceilometer shouldn't be tied to that
[15:28] <zul> MIR as well
[15:29] <Daviey> med_: you rock
[15:29] <jamespage> thanks folks
[15:29] <med_> nod.
[15:30]  * med_ is just now hearing jamespage comments on MRE answers
[15:30] <jamespage> med_, :-)
[15:30]  * med_ would be happy to help
[15:31] <med_> very lagged I gather
[15:31] <med_> Note to self: if you want real time participation, join the hangout.
[15:32] <Daviey> med_: you were invited. :-)
[15:32] <med_> nod.  I just didn't think I had anything to add to this one.... and I didn't really.
[16:00] <SpamapS> o/
[16:01]  * stgraber waves
[16:02] <SpamapS> video died for me :-/
[16:02] <SpamapS> ah came back
[16:04] <Daviey> Hey!
[16:04] <Daviey> Who else wants to be on te inner circle?
[16:05] <Daviey> SpamapS: ?
[16:05] <SpamapS> Daviey: no I'm just watching. Don't have anything to add.
[16:05] <Daviey> Not started yet.. still finding participants
[16:05] <Daviey> ogasawara: Want to join the inner circle?
[16:06] <med_> not live yet
[16:06] <ogasawara> Daviey: I've got another meeting which is about to conflict, so was hoping to watch from the sidelines for now
[16:06] <ogasawara> Daviey: but will be happy to jump in later if needed
[16:07] <ogasawara> Daviey: if you need a kernel rep, lemme see if I can get apw
[16:07] <jamespage> hey channel
[16:07] <jamespage> ping me with any questions for the inner circle
[16:07] <med_> live now
[16:08] <jamespage> or feel free to request to join if you want to be more involved
[16:10] <pitti> don't we already have user namespaces with logind now?
[16:10] <pitti> i. e. sessions already run in a cgroup in saucy
[16:11] <jamespage> pitti, did stgraber answer that question? (stream lag)
[16:12] <pitti> not yet here
[16:12] <jamespage> ok
[16:12] <ogasawara> Daviey: I've got rtg monitoring the session if you've got kernel topics needing discussed
[16:13] <Daviey> ogasawara: I *think* it will be ok, but if he wants to join the inner circle he is more than welcome :)
[16:13] <rtg_> Daviey, I think I'm fine just listening
[16:15] <adam_g> feed just died for me
[16:16] <med_> still alive here
[16:41]  * dannf joins late from another session and searches for the word "arm" in the notes
[16:41] <dannf> any plans for arm/qemu?
[16:42] <dannf> er, qemu-kvm
[16:42] <med_> dannf, you can probably join the hangout if you want
[16:42] <Daviey> Did we cover openvswitch integration to kvm, xen and lxc?
[16:43] <dannf> med_: good point - though i'm not sure how?
[16:43] <med_> see message in another server
[16:43]  * Daviey pm'd dannf 
[16:44] <dannf> thx; waiting till serge is done talking before i switch over
[16:44] <dannf> cortex-a15
[16:44]  * dannf switches
[16:45]  * med_ measures lag
[16:47] <med_> did anyone play with OpenDaylight yet? The code is downloadable....
[16:47] <Daviey> med_: not yet.
[16:47] <med_> not sure how it plays with lxc/libvirt
[16:47] <med_> or --if-- it plays or will play
[16:48] <Daviey> med_: Last week, i couldn't find the source - just the gerrit.. Where is the source?
[16:48] <med_> Daviey, the source is there too.
[16:48] <med_> same place.
[16:48] <Daviey> med_: it ws 404/403/500 error last week i thought...
[16:49] <Daviey> i guess they have fixed it
[16:50] <med_> no, it worked fine with the right credentials
[16:50] <med_> and then someone else posted in an etherpad the write URL rbasak may have.
[16:50] <Daviey> med_: it's authenticated source tree?
[16:50] <Daviey> sounds like semi-open source :)
[16:51] <med_> there was both auth and non-auth
[16:51] <med_> but my checkout is auth:
[16:51] <med_>         url = ssh://medberry@git.opendaylight.org:29418/controller.git
[16:54] <med_> <humor virtualization> what about lguest and the puppies?   http://lguest.ozlabs.org/  </humor virtualization>
[16:55] <med_> is dannf the lxc upstream for debian arm?
[16:55] <dannf> no
[17:51] <Daviey> Anyone here for the Interlock session yet?
[17:58] <ogasawara> Daviey: I'm here, just ping me the hangout when you're ready for me to jump in
[18:05] <Daviey> smoser: are you joining?
[18:07] <zul> feel free to start :)
[18:07] <adam_g> im having trouble joining
[18:07] <smoser> o/
[18:08] <jamespage> http://pad.ubuntu.com/uds-1305-servercloud-s-interlock
[18:08] <rbasak> adam_g: it started late. It's live now.
[18:08] <smoser> * HWE Kernel ramifications on UCA testing
[18:08] <smoser> * ISO default kernels for 12.04.2 server and cloud-images
[18:10] <rbasak> How about the samba4 question? That's more for the desktop team, but I don't think they're present?
[18:11] <rbasak> (though I'm still not sure why we want samba4 default anyway - it's packaged, and on server people install the package they want directly themselves)
[18:12] <rbasak> ARM servers are moving to UEFI, so we'll need MAAS support for that. But as long as d-i works with it, I don't think there will be any issue. I assume secure boot won't be used for now.
[18:13] <rbasak> I'm told that UEFI support is working well with precise-updates on at least one server vendor's hardware.
[18:18] <Daviey> rbasak: ok will ask
[18:19] <med_> UEFI is pretty standard on laptops now.... (not that that means anything w/r/t servers but generally is good for installation story in general)
[18:19] <apw> Daviey, people are still reporting data loss and lack of integity control in btrfs
[18:19] <apw> Daviey, also performance has been poor, plus the space reporting is basically useless
[18:19] <jdstrand> good stuff :)
[18:19] <med_> apw, so you love it?
[18:20] <med_> :)
[18:20] <apw> med_, it is my very favourite
[18:20] <med_> so ceph folk seem to like it but don't use it in real deployments
[18:21] <apw> med_, well with cephs they work on the assumption that if one of their nodes implodes, they can just rebuild it
[18:21] <med_> nod
[18:23] <jamespage> med_, apw: default ceph filesystem is XFS for good reason!
[18:23] <med_> jamespage, nod++
[18:23] <apw> jamespage, heh good :)
[18:25] <med_> El Torito boot image... I heard L32 boot image the first two times....
[18:25]  * med_ needs subtitles
[18:29] <med_> arm64 definitely concerns server
[18:30] <cjwatson> I almost forgot to mention it, oops :)
[18:30] <med_> are there only h/w sims or also s/w siims for Arm64? Ie, can we scale the testing
[18:30] <cjwatson> how do you mean?
[18:30] <cjwatson> I mean what distinction are you drawing there
[18:30] <cjwatson> analogous to qemu-system vs. qemu-user?
[18:30] <med_> well, hardware doesn't scale well to many hands.
[18:31] <rbasak> Did we resolve our questions about the backported enablement stack?
[18:31] <med_> cjwatson, I didn't think there was much arm64 silicon (if any)
[18:31] <cjwatson> we have no hardware yet but some hardware simulators of various levels of performance
[18:31] <cjwatson> qemu is, I hear, expected to give us something in Q3 or so
[18:31] <med_> cjwatson, thanks, that's what I wanted to know. much appreciated.
[18:39] <rtg_> apw, why don't we carry some of these non-binary-blob dkms packages in the ubuntu directory ? That solves the backport issue.
[18:40] <apw> rtg_, a good question.  they have often not wanted to be restricted by our process in updateing them.  I tend to agree that the ones for things like openvswitch which are now completely critical maybe should be spo
[18:41] <s3hh> Will kernel team be willing to consider enabling CONFIG_USER_NS, at least if we add a sysctl to disable unprivileged use of CLONE_NEWUSER by default?
[18:41] <rtg_> apw, those packages are subject to SRU anyways I would think.
[18:43] <rbasak> apw: size of the cloud images?
[18:44] <rbasak> apw: eg. a MAAS deployment needs to download a dozen images maybe, so it multiplies out
[18:44] <rbasak> And then each update of upstream images has to pull them again
[18:44] <med_> It's speed, time, cost.
[18:48]  * med_ dropped due to b/w issues
[18:48]  * rbasak had issues too
[18:48] <rbasak> I got it back though
[18:49] <med_> well, youtube is a lot less demanding than the hangout
[18:49] <Daviey> apw: did you see s3hh question?
[18:49] <med_> and just lost youtube
[18:49] <Daviey> I couldn't easily field that in the session.
[18:50] <med_> or session died
[18:50] <med_> looks like session died
[18:50] <rbasak> Security as an afterthought? Sounds like a number of upstream packages :)
[18:50] <med_> and back...
[18:52] <Daviey> any security questions
[18:52] <Daviey> ?
[18:52] <s3hh> rtg_: ^ ?
[18:52] <s3hh> Daviey: well, there's the qeustion of how apparmor stacking support is looking now, after uds
[18:52] <rtg_> s3hh, looking...
[18:53] <Daviey> s3hh: stacking?
[18:53] <s3hh> Daviey: apparmor profile stacking,
[18:53] <Daviey> ah ok
[18:53] <rbasak> I found privacy extensions odd on my IPv6 server.
[18:53] <s3hh> i.e. load a policy whil still confine dunder another
[18:53] <rbasak> I can't remember the detail - confusing netstat, etc.
[18:54] <rtg_> s3hh, looks like USER_NS depends indirectly on XFS_FS = n
[18:54] <s3hh> rtg_: yeah, obviously only if that is fixed,
[18:54] <s3hh> that, plus a 2-line aufs patch are the only *tchnical* blockers right now
[18:54] <med_> youtube is dropping out, I'll replay later.
[18:54] <s3hh> (test build is in ppa:serge-hallyn/userns-natty working great for saucy)
[18:55] <rtg_> s3hh, send some patches on the k-team list ?
[18:55] <rbasak> Right now, IPv6 on my server is broken and I haven't got round to looking at it. Something to do with the addressing and configuration. It'd be nice if it Just Worked appropiately for server, without extra configuration necessary to make a server behave like a server, IYSWIM.
[18:55] <s3hh> rtg_: once eric gets the xfs uploaded, there'll be no patches.  it'll just be enabling CONFIG_USER_NS,
[18:55] <rbasak> (eg. I can ssh to it find, so ssh is listening on the right socket, but outbound connections don't work)
[18:55] <rbasak> fine
[18:55] <rtg_> s3hh, sounds like thats no problem for saucy
[18:56] <s3hh> rtg_: ok, thaks.  yes, once eric sends the xfs patches for inclusion, there'll be no patches at all.
[18:57] <jamespage> http://reqorts.qa.ubuntu.com/reports/ubuntu-server/cloud-archive/grizzly_versions.html
[18:57] <jamespage> http://reqorts.qa.ubuntu.com/reports/ubuntu-server/cloud-archive/folsom_versions.html